Background The inability to regulate autonomic activity during social interactions is believed to contribute to social and emotional dysregulation in children. Research has employed heart rate variability (HRV) during both socially engaging and socially disengaging dyadic tasks between children and adults to assess this. Methods We conducted a meta‐analysis including evidence from 18 studies comprising 1,544 children who were categorized as either healthy or at risk/diagnosed with psychopathology. Within these groups, we assessed the impact of social engagement and disengagement tasks on HRV. Results Results showed that social engagement tasks left HRV unaltered to a baseline. Social disengagement, however, was associated with decreases in HRV. In a task that included disengagement and then engagement, HRV was reduced during disengagement but was then restored during the reunion phase (engagement). Children at risk or with a diagnosis for psychopathology, however, failed to show any change in HRV during dyadic social interaction tasks. This was despite a social stressor, the Trier Social Stress Test, causing significantly reduced HRV in both groups. Conclusions This meta‐analysis provides support to suggest HRV may provide a worthwhile context specific marker for the effective regulation of dyadic social interactions in children.